Where distributed systems theory meets production scale — and you make it more reliable, performant and cheaper!

Our Open‑Source Roots, Turbocharged in the Cloud

We ❤️ open source. From Linux to PostgreSQL, the best infrastructure stories start in the open and are battle‑hardened by passionate communities. Temporal is no different: the Open Source Server powers thousands of workflows every day.

But when teams outgrow the DIY model—or just prefer shipping features over tuning databases—they move to Temporal Cloud. Thanks to drop‑in migration tooling from our Cloud Global Services team, workflows migrate in hours with zero code changes.

• 🚀 Performance: Customers report 50 % faster feature delivery and p99 latencies in the single‑digit ms range.
• 💰 Cost: One recent migration unlocked $2.25 M in annual savings. (temporal.io)
• 📈 ROI: Forrester’s 2025 TEI study found a 201 % three‑year ROI for Temporal Cloud. (businesswire.com)
• 🛡️ Reliability: Five‑nines durability without the 3 a.m. pager.

Poster‑child migrations like SAP Concur showcase how complex multi‑region data moves become routine with Temporal Cloud workflows. (temporal.io)

And the secret sauce behind the Temporal Cloud? 🥁Cloud Data Store (CDS)! And, yes, we’re hiring! 😄

What is CDS?

Every Temporal workflow is, at its core, a stream of durable state transitions. CDS is the purpose‑built, multi‑tenant system of record that persists those transitions for every customer in Temporal Cloud.

• Massive scale today (2025):
  • 500,000 state transitions per second
  • 450 billion actions per month
  • 8 PB+ of hot workflow history under management
• 99.999% durability for workflow history (five nines—and climbing)
• Single‑digit millisecond p99 latency for the hottest paths

“CDS (Cloud Data Store) is the foundation of Temporal Cloud. It’s the heart of our reliability, scalability, elasticity, performance and reliability (reliability comes first and last!). But it also has capabilities no general-purpose database can match. We both ensure that the company’s current business is always running and provide the basis for future new products.” — Paul Nordstorm, CDS Architect

The Engineering Problem Space

1. Custom home-grown Write-Ahead-Log & cache hierarchy – shard‑aware, re‑entrant, write‑through, globally replicated.
2. Adaptive quorum control – we use technologies that blend Paxos, Raft, and a sprinkle of a secret sauce to keep writes unblocked during region failures. We’re able to recover from a complete availability-zone outage in under 4 minutes! 😄